Editor’s note: Following NuTech Solutions’ acquisition of BiosGroup and the addition of “complexity science” to its repertoire of offerings, Local Tech Wire asked NuTech CEO Matthew Michalewicz to explain what the emerging field is all about, and why entrepreneurs and business executives should be interested.Complexity science strives to uncover the underlying principles and emergent behavior of highly complex systems.
Complex systems are composed of numerous, simultaneously interacting parts or “agents” – from molecules or bacteria in a biological process, to individuals, machines or businesses in an economic setting. Because of its widely applicable analytical capabilities, complexity science can span many disciplines, including physics, biology, and systems theory, to name just a few.
The development of complexity science is not a single technological innovation; it is a shift in scientific approach with the potential to profoundly affect business.
In complex adaptive systems, patterns of behavior emerge that are often unpredictable from an examination of the parts. Birds flock, bees swarm, ants efficiently forage for food – all behaviors that could not be predicted by looking at the individual members of the system. Business examples of emergent behavior include:
- How fads are spawned by the purchase decisions of individual consumers
- How inventory stockpiles build up at specific nodes within a supply network.
The goal of complexity science is to understand these systems: what “rules” govern their behavior, how they adapt to change, learn efficiently and optimize their own patterns of behavior. With this knowledge, a system’s emergent behavior can be understood, modeled and then altered advantageously.
Agent-based modeling
Agent-based modeling is a key technology applicable to a wide variety of business problems, from stock market behavior to supply chain management. It is also known as agent-based or bottom-up simulation. This modeling can be used to analyze and learn how to modify complex systems to benefit the process being examined.
In agent-based computer simulations, software agents represent the components of a business system. Agents’ behaviors are programmed via rules, often established through data mining, so that they realistically depict how business is conducted within that system (e.g., a “truck agent” may be programmed to leave the loading dock only when at 100 percent capacity).
Programming rules can range from simple “if-then” clauses to sophisticated machine algorithms such as neural networks, genetic and evolutionary algorithms and other advanced computational techniques that allow agents to learn and modify their behavior during the simulation, responding to virtual events in the same way that their reality based models would — with diversity and even unpredictable outcomes.
Parameters of the model are set to represent a situation of interest and the model can be run for thousands of iterations. Models are calibrated against historical data to ensure the model is accurately replicating the behavior of the real system and they then can be studied to determine an optimized solution.
Agent-based simulations can provide valuable information about the dynamics of the real-world system that they emulate. As widely varied individual agents interact in the model, the simulation shows how their collective behaviors govern the performance of the entire system – for instance, the emergence of a successful product to meet a particular market need or an optimal pick-up and/or delivery schedule.
What-if scenarios
These simulations are powerful strategic tools for “what-if” scenario analysis: as managers change agent characteristics or “rules,” the impact of the change can be easily seen in the model output. Perhaps most important, the computer can generate strategies that the user might never have considered.
Since the founding of NuTech Solutions, the mission of the company has been to use artificial intelligence technologies through advanced computational techniques, to build prediction and optimization systems that create measurable monetary value. The mission of BiosGroup, the world-leader in applied complexity science, has been very similar, except that BiosGroup has focused on the simulation and modeling aspects of business rather than predication and optimization. Therefore, together, the merged NuTech Solutions and BiosGroup will be able to combine both approaches into one seamless offering, which has the potential of creating value for clients on entirely new and even multiple levels.
Science and business
NuTech Solutions and BiosGroup complement each other in numerous ways. In addition the potent combination of agent-based, optimization, predictive and other advanced computational techniques already described, the combined skill sets of NuTech and BiosGroup will also provide an unprecedented opportunity to both model and analyze the millions of solutions possible in any environment to arrive at optimal solutions regardless of data conditions.
BiosGroup techniques especially thrive in “data-poor” environments while NuTech techniques thrive especially well in “data-rich” environments. Together, NuTech and BiosGroup will be potent in any data environment; the combined resources will be able to “data-farm” as well “data-mine”.
What is important to note is that unlike research institutions or other science-oriented non-profit organizations, the merged NuTech Solutions/BiosGroup is guided by the business needs of the marketplace rather than pure research itself.
The mission of today’s merged NuTech Solutions is to continue to help governments, for-profit and not for profit organizations and other groups obtain measurable benefits in finding optimal solutions: increasing revenue and/or reducing costs through the use of new and innovative technologies.
However, our approach does not develop new and innovative technologies just for the sake of innovation, which is important, because the future development of leading-edge technologies at NuTech Solutions will focus on the solution of real-world problems. The way we blend complexity science, and in particular, agent-based modeling, with our prediction, optimization and other technologies, is guided by the needs of our clients, not the needs of scientific technique.
BiosGroup’s slogan of “Science for Business” will continue to be a primary business goal for NuTech Solutions.
NuTech: www.nutechsolutions.com
